Greedy Orthogonal Matching Pursuit algorithm for sparse signal recovery in compressive sensing

Jia, Li; Wu, Zhaojun; Feng, Hongqi; Wang, Qiang; Liu, Yipeng

doi:10.1109/i2mtc.2014.6860967

Cited by 6 publications

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SLSE has found many applications in modern-day life, e.g., voice assistants [1], smart homes [2], video conferencing [3], acoustic scene analysis [4] and event detection [5], etc. SLSE methods fall into two categories: the one-stage methods [6], [7] and the two-stage methods [8], [9], [10]. One-stage methods conduct source localization and signal extraction in one step, whereas two-stage methods localize the sources first and separate the signals emitted from the sources in a subsequent step based on the information gained in the first stage.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Source Localization and Signal Extraction Using a Proximal Gradient-based Compressed Sensing Approach

Tao¹,

Chen²,

Hsu³

et al. 2022

JAMC

View full text Add to dashboard Cite

This paper presents a computationally efficient algorithm for multiple source localization and signal extraction (SLSE). Posed as an underdetermined system, a novel compressed sensing (CS) algorithm is proposed to address SLSE problems in one stage. A Least Absolute and Selection Operator (LASSO) problem is first formulated and solved jointly for the source locations and signal amplitudes. A computationally efficient and noise-resilient algorithm is developed on the basis of the complex Proximal Gradient (Proxgrad) method. It follows that the nonzero entries of the optimal solutions give rise to the amplitudes and directions of sound sources. To further enhance the separation quality, soft thresholds based on W-disjoint orthogonality is exploited. Experiments are conducted to compare the proposed SLSE method with several baselines in terms of localization and separation metrics. The results showed that the proposed LASSO-Proxgrad algorithm yielded superior localization and signal extraction performance with the minimal processing time compared to the baselines.

show abstract

Section: Introductionmentioning

confidence: 99%

“…The one-stage method is based on the notion of Compressed Sensing (CS), also known as Sparse Coding [6], [7]. Although CS can be applied to many problems, relatively few application examples in acoustic array signal processing problems are found in literature.…”

Section: Introductionmentioning

confidence: 99%